5 Detailed policy targets
The analysis above was conducted on the basis that policy is defined over three dimensions, each dimension representing the aggregate economic welfare impact either of distortion taxation, time-consistency or agency cost. This was a simplification to make the analysis tractable. In contrast, the background discussion on objectives and targets (Section 2) identified 15 dimensions or targets over which policies may be defined (refer also to Table 6 below). The purpose of this section is to refine the three candidate policies in terms of how they would apply to the 15 dimensions. The status quo policy is recorded for comparison purposes.
Non-controversial targets
It is sometimes the case that the various targets associated with an objective have quite different risk properties. It is of interest to identify whether any particular targets would be particularly low risk and non-controversial. One test of whether a target would be non-controversial is whether it is supported by some objectives and contradicted by none. A policy target meeting this test could be adopted as part of the “baseline” applicable to all policy options.
Three of the 15 dimensions are considered non-controversial:
- Tax smoothing over the cycle (refer row 1 in Table 6 below). Full smoothing of tax rates over the economic cycle is supported by tax distortion considerations. It has also explicitly or implicitly been part of the status quo policy for many years. Full tax smoothing over the cycle is potentially contradicted by the agency-costs objective, but in practice is unlikely to be a significant issue provided the amplitude and duration of cycles are not too large or long.
- Upper bound on share of assets held (row 8). Both large player and agency cost considerations imply an upper bound on the proportion of any asset held by the Crown.[31] Potentially this could be contrary to the objective of minimising distortionary tax if empirical correlations indicated a large weight on NZ assets, but this appears unlikely.
- Lower bound on quantities of benchmark bonds (row 11). Placing a lower bound on the quantities of benchmark bonds on issue is aimed at avoiding high liquidity premia. For reasons similar to those above, the proposal is unlikely to be contrary to other objectives.
These three targets have been included in all policy options (refer Table 6, rows 1, 8 and 11). This means, for example, that the low risk policy (L,H,H), despite placing low weight on distortionary taxation, incorporates full tax-smoothing over the economic cycle.
Basis for specifying remaining targets
For the purposes of this paper, it is assumed that future work will establish a plausible range of values for each policy target. It is assumed that, for each target, one end of the range of values would represent a strong or conservative stance while the other end of the range would represent a weaker or more ambivalent stance. The proposal is to specify the policy targets as follows:[32]
1. A policy that assigns high weight to an objective (that is not countered by high weights on other dimensions) would be represented by choosing values for the relevant targets at the strong or conservative end of the range, e.g. “strong” upper bound on total debt;
2. A policy that assigns low weight to an objective would be represented by choosing values for the relevant targets at the weak end of the range or specifying no target at all; and
3. Where the weights assigned by a policy imply conflicting assignments, the values for the relevant targets would be chosen at the middle or weak end of the range as appropriate.
The remainder of this section discusses the application of this approach to the three candidate policies. Results are summarised in Table 6 below.
Low risk policy (L,H,H)
In the low risk policy, the high weight on agency cost would translate into strong upper bounds on the Operating Balance and fungible assets. There are no countervailing forces as distortionary taxation receives low weight.
The debt target is more complicated: The high weight on agency cost implies high debt levels (i.e. a strong lower bound set at the high end of range), while the high weight on time-consistency implies low debt levels (i.e. a strong upper bound set at the low end of range). Whether these assignments are conflicting is an empirical matter, and may change over time. If they do not conflict then the policy implication is that the level of debt would be maintained within a collar defined by the upper and lower debt targets. If they conflict then it is proposed that the value of the debt target would be the midpoint of the upper and lower bounds.
| Targets | Status quo [33] | Low risk policy (L,H,H) | Medium risk policy (H,H,H) | High risk policy (H,L,L) |
|---|---|---|---|---|
| Risk/return properties | ||||
| Tax-smoothing over cycle | Full | Full | Full | Full |
| Deterministic smoothing | Partial | - | Partial | Full |
| Diversifiable risk | - | - | Minimise | Minimise |
Systematic risk | - | - | Partial hedge | Minimise[34] |
| Net worth | ||||
| Bound on cyclically-adjusted OB | - | Strong upper | Weak upper | - |
Positive CNW buffer | - | - | Small[35] | Large |
| Financial asset structure | ||||
| Bound on fungible assets | - | Strong upper | Weak upper | - |
Bound on share of asset held | Upper | Upper | Upper | Upper |
| Debt structure | ||||
| Bound on total gross/net debt | ≤ 30% | Collar[36] | Collar | Strong upper[38] |
| Bound on debt maturing | < $3.5bpa | Upper | Upper | - |
| Bound on benchmark quantities | ≥ $3b | Lower | Lower | Lower |
| Bound on average maturity | - | Strong lower[37] | Strong lower | - |
| Bound on inflation-indexed debt | - | Strong lower | Strong lower | - |
| Bound on net foreign-currency debt | Zero | Strong lower | Strong lower | - |
Bound on floating rate instruments | 20-30% | - | - | - |
Note: The table is read from right-to-left in the sense that, for example, “Full” in the first row of status quo column is read as “full tax-smoothing over the economic cycle. “Strong upper” means “strong upper bound on ….”, implying that the target is expressed as an upper bound on the relevant variable and “strong” means the value assigned would be near the conservative end of plausible values to provide a strong constraint against jeopardising the relevant objective. “Weak” means the value assigned would be near the opposite end of plausible values to provide a weaker constraint against jeopardising the relevant objective. “Collar” means there is both an upper and lower bound on the relevant variable.
Notes
- [31]The upper bound might also be expressed as a nominal dollar limit rather than only in proportional terms.
- [32]The proposal assumes that the main interest is to consider the long-term direction of Crown financial policy. If instead the main interest is to conduct marginal analysis of the next step along a transition to the long-term policy target then the values assigned to the policy targets would be small changes from the status quo policy. The remainder of analysis would remain largely unchanged.
- [33]Under the status quo the policy targets relating to the $3.5 billion upper bound on debt maturing, $3 billion lower bound on benchmark debt and 20-30% target for floating rate instruments are specified for NZ dollar denominated debt. The upper bound on debt maturing excludes Treasury Bills. The target of zero net foreign-currency debt applies where the definition of net foreign-currency debt is gross foreign-currency debt net of foreign exchange reserves and NZDMO-managed foreign-currency assets.
- [34]Systematic risk should not be minimised in the case where citizens are non-responsive and the government objective is to minimise expected deadweight losses.
- [35]Positive balance applies if and only if negative correlation between tax rates and consumption, otherwise zero balance. Same condition applies to other policy options where relevant.
- [36]Time-consistency implies an upper bound while agency cost implies a lower bound. If not conflicting, the two bounds would form a collar. If the bounds are conflicting (in the sense that the lower bound exceeds the upper bound) then the target would be set at the midpoint between the two bounds. This applies also to the ‘medium risk’ column.
- [37]Lower bound is contingent on breach of the threshold upper bound for total debt. The same condition applies to the lower bounds on inflation-indexed and foreign-currency debt under all policy options where relevant.
- [38]Strong upper bound is implied by the distortionary tax objective to the extent that “unjustified risk premia” occur at high debt levels.
